Just taught my Halloween special lecture in computer security and am reminded about the ease of fooling yourself. The lecture itself is about some mis-leading statistics and a bit of slight of hand. The use and more importantly the design of control experiments is critical in science to help avoid fooling yourself.
Slight of hand itself is a good example of the fooling yourself. You can’t mislead your audience unless you, sort of, mislead yourself. When doing the “french drop” or the pass vanish – you have to look at where the coin or ball should be and not where it is. The magic isn’t convincing unless you look like you”re convinced too. Smoothness helps, but the audience looks where you direct them.
In science, getting the answer you want and expect is always troubling. If you aren’t wrong most of the time, then you aren’t working at the bleeding edge. You need to moral courage to keep at it. This is one of the hardest things to teach, but one of the most critical things that my students need to learn.